Joint SCSP-LROM: A novel approach to detect Cerebrovascular Anomalies from EEG signals

TL;DR

This paper introduces the SCSP-LROM method, combining sparsity and low-rank optimization, to improve detection of cerebrovascular anomalies from EEG signals with high accuracy.

Contribution

It presents a novel joint sparsity and low-rank optimization approach for EEG analysis, enhancing anomaly detection accuracy over existing methods.

Findings

Achieved 96.3% overall accuracy in detecting tumours and lesions.

Developed a new optimization model based on compressed sensing principles.

Enhanced EEG channel selection for better prediction in brain computer interfaces.

Abstract

It has always been a big challenge to identify subtle changes in Electroencephalogram (EEG) signals. Minor differences often lead to vital decisions, for example, which grade a certain tumour belong to or whether a haemorrhage can result in benign blood clots or cancerous ones. In recent studies on brain computer interfaces (BCIs), one of the biggest challenges is recovering maximum information for realistic predictions. In order to choose EEG channels with highest accuracy, a novel notion of including sparsity in a modified common spatial pattern (CSP) algorithm is introduced here. Being influenced by the existing concept of compressed sensing, an optimization model is also developed alongside to recover the cosparse signal and retain maximum information. The state-of-the-art Joint Sparsity Induced Modified Common Spatial Pattern Algorithm and Low Rank Optimization Model (SCSP-LROM) developed here is capable of identifying and describing tumours and lesions in great detail at an overall accuracy of 96.3%.